Getting Started

Prerequisites

You need Claude Code installed. The skills run as structured prompts inside Claude Code sessions.

Installation

Option 1: Run from the repo

git clone https://github.com/benjam3n/reasoningtool.git
cd reasoningtool/claude-code-plugin
claude

Claude reads the CLAUDE.md file and all 207 skills become available.

Option 2: Copy into your project

cp -r path/to/reasoningtool/claude-code-plugin/skills your-project/
cp path/to/reasoningtool/claude-code-plugin/CLAUDE.md your-project/
cd your-project
claude

Your first session

In a Claude Code session, type a skill name followed by your input:

/araw should I change careers

This runs an ARAW analysis: Claude takes the claim "I should change careers," assumes it's right (what follows?), assumes it's wrong (what are the alternatives?), and recurses on the most important sub-claims. The output is a tree of tested claims with crux points and action items.

More examples

# Map all options for a product launch
/space_enumeration what are all the ways I could approach this product launch

# Surface hidden assumptions
/assumption_extraction we need to hire 3 more engineers by Q2

# Structured comparison
/comparison React vs Svelte for this project

# Find the real cause
/root_cause_analysis our deploy pipeline keeps failing

# Full alternating search (exploration + testing + edge cases + validation)
/uaua 4x what's the best approach to learning math

Depth scaling

ARAW and UAUA support depth scaling. Higher depth = more thorough exploration:

Depth	What you get	Use for
`/araw`	Quick analysis, 3–4 levels deep	Simple claims, low stakes
`/araw 2x`	Deeper, 4–5 levels, more claims	Moderate complexity
`/araw 4x`	Thorough, 5–6 levels, edge cases	Important decisions
`/araw 8x`	Exhaustive, 6–8 levels	High-stakes analysis
`/uaua 4x`	Full alternating search at 4x depth	Complex problems needing both breadth and depth

Python tools

For programmatic exploration with persistence and visualization:

pip install openai pyyaml
export OPENAI_API_KEY="sk-your-key"
cd src/araw

# Auto-expand a claim into a branching tree
python auto_expand_llm.py --db career.db --seed "Should I start a business?"

# View in browser
python visualize.py --db career.db --serve

See Tools → for the full documentation.

What to try next

Pick a decision you're facing and run /araw on it
Browse the 207 skills to see what's available
Read Two Kinds of Search for the philosophy behind the alternation pattern
Try the Python tools for persistent, visual exploration